You can also just do "uv run", and it will automatically sync as necessary!
I think that's one of the things that uv does great and often goes unnoticed: It's hard to use wrong for a layman. Forgetting to run the equivalent sync command in most other Python package mangers could have resulted in a myriad of different bugs caused by out of sync packages, leading to hours of wasted debugging.
> In a uv project I can just git clone, cd, uv sync and I am ready to go. Nothing in the python ecosystem offered that before.
`poetry install` does the same thing except for downloading Python.
I transitioned from Poetry to uv. Other than being mildly faster and providing workspaces, uv doesn't really change at all how one interacts with and manages a project. And in fact, uv just adds additional quirks and gotchas and limitations, including storing authorization keys in plain text.
uv also duplicated things in the Python ecosystem but then didn't implement the same feature sets, so you often end up needing to call the existing tool, just like you did in Poetry.
I suspect, that lots of people hyping uv didn't have any proper reproducible project setups beforehand and didn't have to find solutions to that before uv. If they had, they would know about Poetry and how to use it.
I am of the same opinion, that uv didn't actually change all that much, if you have a sane project setup. I don't do anything with it, that I didn't do before with Poetry and before that with Pipenv. What I need is a way of installing dependencies, checked against a lock file, reliably. Poetry does that. If uv disappeared today, I would go back to Poetry and wouldn't miss all that much.
But there are the people who want to write one-off scripts with loosely defined dependencies inlined into the scripts. I don't do that. I would just copy my Makefile and pyproject.toml file from another project and adjust it and then run the script just as well with locked dependencies. Why do anything less. If I want to be really sloppy, I can just pip install.
I tried really hard to do things the right way before uv, but it was just a major PITA. Poetry in particular looked really exciting when I first discovered it, but I eventually just got bogged down with the speed and its inability to resolve dependencies even for recent versions of pretty mainstream packages. It didn't 'just work' once you got beyond toy projects.
I admit I hadn't really realised astral was a private company before though. What is their roadmap for monetising all this work?
Hm, that's strange. I have used Pipenv and Poetry years ago, to reproducibly build data science environments that people used to learn about data science and data analysis. I don't think you can call that "toy projects". I guess you just had bad luck with some specific dependencies. What I do remember though is, that Poetry had far fewer problems than Pipenv had. Pipenv was still a fight, though I won every time one way or another.
Yea, it's interesting. My comment is negative in votes, but I went back to it to see. Everything I said is factual, so I don't know. People are endlessly impressed by uv's whizbang speed, it seems. In reality, it's _yet another tool_ trying to reinvent the Python ecosystem.
About half the t/s of an online provider, heat, noise, power consumption and on top of that you only get 75% of the intelligence? Sounds like a waste of time to me.
Half the token/s - sure, no way around beating big ass data centers.
Heat, noise and power consumption are actually quite ok - i'm running 2x3090s, capped at 250W and i haven't seen heat and noise being an issue, sitting right next to it. Might be helping that it's getting colder.
75% of the intelligence with 2x3090s, with a 5090 if you have it it's around 90% and these are first day builds too, typically people hone in and improve these all the time.
It's true local doesn't make sense value wise against online, but if you have the hardware, having "unlimited" tokens with no 5h limit, no weekly limit, is super nice.
As luck would have it, you can just pay the provider and use their service. Others may want to run locally for privacy reasons or just for the fun of it. I think at-home AI is incredibly awesome and I applaud the enthusiasts pushing it forward. Just because that's where it's at today doesn't mean it's the best we can do.
Plus, you know, this thread was about running AI on "lame consumer hardware". There are more power efficient options available.
I use only providers because I don't have the hardware, but if, and that's a big if, the only moat left for humans is our ability to steer and make bigger picture decisions with our personality, I think it makes sense not to want to send all that conversation data to a 3rd party.
This 100%. My current workflow is switching between CC and VS Code to review code, and it's extremely clunky. There is still not a single harness/ADE that can do everything correctly in one tool.
Between VSCode and AI, all the engineers I work with have stopped using JetBrains products.
I don't think the problem is AI itself; they can have AI too, AI as capable as something like Cursor, for example. The issue is that their product is bloated and slow, and much of the technological leadership they once had has been overtaken by VSCode.
Jetbrains IDEs are far from bloated and slow especially compared to electron and chromium based software. Besides - they are still growing in revenue so the financial loss probably isn't due to users jumping ship.
For Android development I still prefer Android Studio (IntelliJ based). Maybe just a habit because I don't really write code there anymore. For everything else usually VS Code.
You could. But now you are trusting an entirely new toolchain that can build the Rust-based linker, even if it is statically linked. This toolchain is much larger than a comparatively small and much more easily understood C bootstrap toolchain.
This effectively triples or quadruples (maybe even more) the amount of code you need to trust for the cold bootstrap.
Too many years at Google where the two terms are most often used interchangeably; which is imprecise, but most of the time it doesn't matter. Forgive me for being imprecise here.
The fact remains that rust adds yet another toolchain to the boot process that needs trust and verification.
What output style have you found to actually fix Opus 5's grating prose then? I find it leans hard into its preferred grammatical structures and rote sayings no matter what I include in the output style.
reply